Abstract. In the study the method of determination of the experiment size is provided so that the optimum reasoning could be applied on the basis of experiment results. The parameter estimation is examined here at the set absolute or relative error as well as hypothesis testing at pre-set (high) power. The considerations contained in the paper refer to success probability in binominal distribution. Such a distribution frequently occurs in the research in agricultural engineering.
Key words: Test power, sample size, binominal distribution, success probability.